Strategic Flexibility in R&D: how to use project selection to prepare for an unpredictable future

Research-Technology Management, May-June, 2004 by Michael E. Raynor, Ximena Leroux

Project selection is a critical lever in increasing the effectiveness of the research function in technology-driven companies. As a consequence, there is no shortage of analytical frameworks to help guide R&D managers in making project selection choices. From relatively simple qualitative algorithms to complex financial return models, the aim is the same, straightforward and maddeningly difficult to achieve: to pick the projects that will yield the highest returns.

Existing frameworks fall into two broad categories: "quantitative" and "strategic." Based on our work with a leading R&D organization, we believe that both approaches, while helpful, are insufficient. Specifically, in the quest to quantify the impact of individual projects, most existing valuation tools run far too high a risk of turning a blind eye to portfolio deficiencies and project interdependencies. The more well-known approaches that seek to align R&D efforts with competitive strategies are an improvement in this regard, but perhaps somewhat ironically, they may increase strategic risk and tend not to cope well with the necessarily dynamic nature of the strategies they seek to support.

In this article we show how "Jupiter Systems," a disguised leading industrial research organization, has applied a new framework, Strategic Flexibility, to overcome these limitations. We also offer some generalizable principles extracted from Jupiter's experiences so that other R&D managers can begin to think about how Strategic Flexibility might inform their own project selection decisions.

Limits of Traditional Methods

In a world of limited resources, R&D organizations have to choose which projects to fund or continue to fund from a set of proposed and ongoing projects. There are sophisticated financial valuation methods for most large corporate investments, such as NPV and option value, and some of these tools have been applied successfully to R&D projects (1,2,3). However, there is a class of research projects for which these tools are inappropriate because the technologies under investigation are far enough removed from commercial products that revenue and cost projections are not feasible.

A popular solution has been to use heuristics that allow managers to rank-order research projects (4,5,6). These heuristics assess the possible impact of a particular project by taking into account estimates of its market potential, science and technology relevance, alignment with corporate strategy, and other qualitative factors. Typically, a weighted average of the different measures is calculated, yielding a ranking of the projects under consideration based on their relative values. After the ranking is complete, resource allocation can proceed, starting with those projects with the highest grade. Note that the guidance these approaches provide is not nearly as precise as more quantitative measures: knowing that project A is more valuable than project B does not provide much insight into how much project A is actually worth. But what they lack in precision, these approaches often make up in accuracy.

Such rank-ordering heuristics can be a powerful tool for project selection when applied carefully to a well-selected set of projects. However, they also have limitations. We can understand the limitations by examining the basic elements of these heuristics: a project set, criteria for evaluation, a ranking and consequent resource allocation. Problems in any of these elements will result in problems with the project selection decision.

First, all rank-ordering heuristics start with a project set from which some projects will be selected for funding and some may be turned down. It is generally taken for granted that all the projects the organization needs to undertake are included in that set. However, if for any reason some important projects are missing from the project set, these heuristics cannot uncover that deficiency, so that the resulting portfolio will also be deficient. Incomplete project sets may be the result of excessive focus on strategic alignment, discussed below. Similarly, any deficiencies of any given project can be exacerbated by such focus.

The second element of rank-ordering heuristics is the criterion for evaluation. Projects are usually evaluated on multiple dimensions, such as potential market impact, science and technology eminence, probability of success, and alignment with corporate and business unit strategy (7). There are two potential problems with these evaluation criteria: an excessive focus on strategic alignment, and project interdependency.

At first blush, insisting on strategic alignment between research efforts and a corporation's strategy seems uncontroversial. However, when strategic alignment becomes a defining factor of project attractiveness in qualitative algorithms, it may constrain the R&D function in ways that actually undermine future competitiveness. Giving priority to those projects that support the explicit corporate strategy (and it is very difficult to support a corporate strategy that has not yet been articulated--but more on this below) ignores one of the most important functions of R&D: to create the technologies on which the corporation's future strategy might be based. Because R&D is one of the few functions that is in a position to create the right set of choices for a company's future, relying on the current corporate strategy as a determining element of project funding may optimize for the current strategy at the expense of the company's future choices.